[ripperjack]

Hi all,

I conducted some DNS work on full channel flow, I wanted to save some computational resource so I jumped from full channel to open channel and I tried to re-produce all the flow condition in a open channel but something is not what I wanted:

For the open channel case, I used slip boundary condition for U at the upper wall (symmetric plane in full channel), I found that the v (normal component) fluctuations, i.e., v_rms, is 0 at the upper wall, but I check the result for the full channel flow, v_rms and is NOT 0. So is there better boundary condition for the upper wall in a open channel which will not enforce v_rms to be 0?

You guys can help me please? Thanks in advanced.

Regards,